AI data centers, a political issue this election year
AI data centers are emerging as an unexpected flashpoint as voters prepare for the upcoming general elections. Across the nation, local communities are questioning whether these facilities’ enormous demands for electricity, water, and land are compatible with available resources and long-term sustainability.
In areas where the expansion of AI data centers has contributed to rising utility bills, the debate has moved beyond technology and infrastructure. It has become a kitchen-table issue for voters who they say are already struggling to make ends meet and do not want to pay additional expenses for AI data centers.
Bipartisan Issue
While there are deep dividing issues between Republicans and Democrats, concerns over AI data centers have largely been bipartisan.
Independent Sen. Bernie Sanders, the first federal legislator to call for a national moratorium on AI data center construction, said “These facilities consume more electricity than entire cities while paying a fraction of the taxes. Working families subsidize corporate AI development through higher electricity bills and degraded grid reliability,” he said. “Frankly, I think you’ve got to slow this process down.”
Republican Gov Ron DeSantis of Florida has publicly opposed hyperscale data center expansion in Florida and proposed an AI Bill of Rights allowing communities to block them. “We have a limited grid. You do not have enough grid capacity in the United States to do what they’re trying to do. Do you want a hyperscale data center in The Villages? Yes or no? I think most people would say they don’t want it,” he said.
The debate has intensified as scientists warned that AI data centers could double national electricity use within a few years and consume millions of gallons of water each day for cooling.
Candidates across the political spectrum now find themselves navigating a complex landscape: voters who welcome technological innovation but fear its local impacts. The issue has become a litmus test for how leaders balance economic development with resource protection, transparency, and community consent. As election season unfolds, AI data centers have become a symbol of the broader question confronting voters: who benefits from the AI revolution, and who bears its costs.
National protests
Opposition to the rapid expansion of AI data centers is drawing national attention. On July 18, HumansFirst coordinated 142 demonstrations across 42 states, according to Reuters. Turnout varied by location, but campaigns across the country have packed public hearings, gathered petition signatures and helped delay or block some developments, signaling that community resistance has become a significant obstacle to the AI infrastructure boom.
In California’s Imperial Valley, residents protested a planned 330-megawatt complex after its developer sued for access to about 260 million gallons a year of Colorado River water, intensifying concerns about diverting scarce supplies from farms and communities.
National energy and water consumption of AI data centers
U.S. data centers are poised to claim a rapidly growing share of the nation’s electricity supply. McKinsey & Company estimates that consumption could rise from 147 terawatt-hours in 2023 to 606 terawatt-hours by 2030, increasing from 3.7% to 11.7% of total U.S. power demand.
The power requirements of individual projects are also escalating. Deloitte reports that the largest AI data centers now being planned could require as much as 2 gigawatts, while early-stage mega-campuses could consume 5 gigawatts — enough to serve about five million homes. At that scale, a single facility can draw electricity comparable to a small city, increasing pressure on utilities, power grids and communities.
AI data centers also place demands on water supplies both directly, through cooling, and indirectly, through the generation of electricity. A 2026 Harvard analysis of 472 large U.S. facilities estimated an annual operational water footprint of about 300 billion liters, or 79 billion gallons, with roughly three-quarters tied to electricity production and the remainder used mainly for on-site cooling. For perspective, 79 billion gallons is enough to fill roughly 120,000 Olympic-size swimming pools. It also equals about 216 million gallons every day, averaged over a year.
What the scientific community says about AI data centers
Scientists examining the rapid expansion of AI data centers warn that the industry’s environmental footprint extends well beyond electricity use. Researchers also caution that the full cost of scaling AI often remains obscured because company disclosures and conventional assessments may not capture the combined effects of energy, water, land use and supporting infrastructure.
A Lawrence Berkeley National Laboratory report led by Arman Shehabi, one of the leading U.S. researchers on data center energy use, found that U.S. Data-Center power use could double or triple by 2028. He said, “Data centers are one of the fastest-growing electricity loads in the world… Their energy use could double in just a few years.”
Emma Strubell, Carnegie Mellon University, a researcher on AI model training estimated that training one especially compute-intensive language model could produce nearly five times the lifetime emissions of an average American car.
A study led by Shaolei Ren, University of California, Riverside, the leading researcher on AI water consumption, estimated that training GPT-3 in Microsoft’s U.S. data centers could “directly evaporate 700,000 liters of clean freshwater,” and concluded that AI’s water footprint “has largely remained under the radar.”
Kate Crawford, Researcher at USC & Microsoft Research and author of Atlas of AI, widely cited on environmental harms has described AI as an extractive system supported by minerals, energy, labor and vast physical infrastructure, arguing that its planetary costs are obscured by the technology’s seemingly immaterial form. She said, “The environmental costs of AI systems are substantial — from water use to carbon emissions — yet they remain largely hidden. AI is built on the extraction of energy, labor, and natural resources. Data centers consume enormous amounts of electricity and water.”
Dr. Jesse Jenkins, Princeton University Energy Systems Researcher, expert on grid reliability and large electrical loads, told the New York Times, “The growth of data centers is putting real pressure on electric grids… AI accelerates that demand dramatically.”
Hawaii data centers
Hawaii currently has 10 operational data centers, with 0 under construction and 0 planned, according to the most recent statewide infrastructure tracking. These facilities represent 27 MW of reported capacity, though only half publish capacity figures, meaning the true total is likely somewhat higher, and could be as high as over 20, depending on the directory and its counting method.
Data center map lists 9 facilities across Honolulu, Kapolei, and Waimea. There are others on federal property. Although totals vary slightly because databases use different inclusion criteria, these are primarily colocation, telecommunications, government, and university facilities — not large hyperscale AI campuses.” National counts also confirm Hawaii as one of the states with the fewest data centers, tied near the bottom of the U.S. list.
Hawaii does not yet have hyperscale AI data centers, so there is no published, Hawaii-specific breakdown of electricity use, renewable energy consumption, or water usage for AI data centers. AI Data centers in Hawaii are still mostly conceptual, and emerging designs emphasize on-site renewable generation rather than grid dependence.
Experts say Hawaii’s grid cannot support a 100–500 MW facility without major upgrades, so any Hawaii-based AI center would almost certainly rely on on-site generation (solar, hydrogen fuel cells, battery storage) with the grid as a secondary backup. Hawaii mandates 100% renewable energy by 2045 (Oahu), 100% renewable by 2035 (neighbor islands).
What Hawaii’s water usage would look like if Hawaii built a 20–50 MW AI data center — traditional evaporative cooling could require millions of gallons per day. Hawaii’s water constraints make this unsustainable, data shows.
Hawaii lawmakers on AI data centers
Hawaii lawmakers are calling for closer scrutiny of large data centers as artificial intelligence and cloud computing drive demand for energy-intensive facilities. In 2026, Senate Resolution 90 and House Concurrent Resolution 206, House Draft 1, asked the Hawaii State Energy Office to convene a working group to study effects on electric utilities, ratepayers, natural resources and climate goals. The measures warn that major facilities could require costly grid upgrades, increase outage risks and consume tens of millions — or, at the largest scale, billions — of gallons of water each year. They also emphasize that Hawaii’s isolated island grids are especially vulnerable to large new loads and say safeguards are needed to keep infrastructure costs from shifting to households and small businesses.
Hawaii’s Filipino community have concerns
Respondents broadly agreed that large AI data centers would require close scrutiny because of their potential demands on the islands’ limited water and electricity supplies.
Rhoda Balmori called for more public understanding of the technology and said developers should bear added energy costs rather than pass them to residents. “When the issue or topic involves water and energy, it is only natural for us to show interest and concern especially regarding the (would be) consumption levels of AI data centers, which are not yet fully clear to everyone. However, we may first need to understand what AI data centers are and consider the benefits they can offer us.”
Balmori, Honolulu, added “it is only right for the companies building AI data centers to take responsibility and shoulder the additional energy-related costs. It would be unfair for us — and especially for those not relying on or use modern technology — to shoulder additional utility bill increases caused by the AI data centers built.
“The benefits AI can provide may outweigh the costs and demands required to build it, but thorough study, careful control, and rigorous oversight by the involved agencies and the companies building it are extremely necessary.” Balmori said she trusts that the government, various concerned agencies, scientists, researchers, and industry groups will subject this argument to a thorough study and that she eagerly waits for any updates they’ll share.
Rose Cruz Churma similarly favored environmental review and urged companies to explore solar, wind and ocean-water cooling. “It’s been shown that there is a toll on the environment. As such, data centers should be subject to an environmental assessment — and if needed, an environmental impact statement for proponents to show what environmental impacts are expected, and their solutions to it to minimize impact. The State laws already mandate this for certain projects, such as those using public funds or on public lands. Perhaps the laws should be expanded to include creation of data centers,” she said.
On whether the benefits of AI are worth the demands that’s needed to power it, Churma commented “Whether we like it or not, AI is already embedded in our daily lives. We are already reaping the benefits (as well as the downsides) even before we realize the full impact of AI on the environment and our humanity. Some jurisdictions have imposed moratoriums on the creation of more data centers. Perhaps Hawaii should do the same and allow our scientists and experts to assess its full impact on our fragile ecosystem?”
Churma said even if a national moratorium (freezing the building of AI data centers) occurred, “we have very little control on what goes on within military installations. The nature of military roles requires them to be secretive and their activities to be strictly confidential. Also, a major engine of our economy is military spending.
“I am really skeptical that a moratorium will include our military facilities since the geopolitical scenario requires them to be at par — technology-wise, with other nations. And unfortunately, it is the military that will need the use of data centers here in Hawaii. Despite all these, each of us should try to learn as much as possible on the impacts of data centers and AI, instead of burying our heads in the sand,” Churma said.
Carlota Ader warned that data centers could strain water supplies, raise living costs and threaten privacy but she doesn’t support a moratorium and favors data centers as a boost to the local economy. “The typical Hawaii resident does not know much about data centers and their impacts on the community. Having said that, it is the responsibility of legislative bodies that approve building and land use permits and other state agencies such as the Public Utilities Commission to carefully vet these large-scale ventures and the impacts on our state,” Ader said.
“For decades, public and private sector leaders have discussed and encouraged efforts to diversify Hawaii’s tourism-dependent economy and to promote sectors such as the film industry, sports and technology to provide a financial boost and as well as jobs. AI data centers are ideally-suited to do just this—provided the process is done correctly and most importantly, that taxpayers’ pocketbooks are unaffected,” she added.
On government’s role and AI companies’ responsibilities, Ader commented, “it is important for these multi-billion dollar companies to pay their fair share of the expenses associated with AI data centers, including electricity, water and other utility costs. The state has no business offering tax incentives and other subsidies to these wealthy corporations. In exchange for land development and other building permits, it is only fair to assess these companies’ exaction fees to pay for infrastructure improvements such as sidewalks, improved roads, playgrounds, parks and other quality of life necessities for the communities that will be impacted.”
Ader believes Hawaii is the ideal place for data centers, especially on the Big Island where there are vast tracts of available land near the coastline.
Memito Ablan said large AI data centers could threaten Hawaii’s limited freshwater supply unless they use alternatives such as ocean water and argued that developers — not residents — should pay for the facilities’ energy and maintenance costs. “AI companies should be required to foot the bill for the on-going cost and maintenance of data centers. The financial burden should not have to fall on the people of Hawaii. Taxpayers cannot afford it, since we already face a high cost of living and many of us are living from paycheck-to-paycheck.”
On a possible national moratorium, Ablan said he supports it. “Data centers should not be built until the public is assured that these facilities won’t harm the environment and won’t increase the cost of living. They are using precious resources.”
About Author
+ There are no comments
Add yours